人工智能的四皇后问题，回溯法——在约束条件下先序遍历，并在遍历过程中剪去那些不满足条件的分支。

使用回溯算法求解的问题特征，求解问题要分为若干步，且每一步都有几种可能的选择，而且往往在某个选择不成功时需要回头再试另外一种选择，如果到达求解目标则每一步的选择构成了问题的解，如果回头到第一步且没有新的选择则问题求解失败。 在回溯策略中，也可以通过引入一些与问题相关的信息来加快搜索解的速度。对于皇后问题来说，由于每一行、每一列和每一个对角线，都只能放一个皇后，当一个皇后放到棋盘上后，不管它放在棋盘的什么位置，它所影响的行和列方向上的棋盘位置是固定的，因此在行、列方面没有什么信息可以利用。但在不同的位置，在对角线方向所影响的棋盘位置数则是不同的。可以想象，如果把一个皇后放在棋盘的某个位置后，它所影响的棋盘位置数少，那么给以后放皇后留下的余地就太大，找到解的可能性也大；反之留有余地就小，找到解的可能性也小。-Four Queens problem of artificial intelligence, backtracking- under constraints preorder, and cut branches that do not satisfy the conditions during the traversal. Characterized in use of backtracking algorithm for solving the problem should be divided into several steps, and each step has several possible choices, and often unsuccessful when you need to go back and try a selection Alternatively, if the target is to reach the solved every step of the solution of the problem constitutes a choice, go back to the first step and if no new options to solve the problem failed. In backtracking strategy, also by introducing some of the problems associated with information to speed up the search for a solution. For queens problem, because each row, each column and each diagonal, can only put a queen, a queen when placed on the board, no matter what position it on the board, and it affects the rows board position in the column direction is fixed, row, column, there was no information about what